Release DepthSplat on Hugging Face

[NielsRogge]

Hello @haofeixu 🤗

I'm Niels and work as part of the open-source team at Hugging Face. I discovered your work as it was trending on PapersWithCode and indexed the paper page here: https://huggingface.co/papers/2410.13862. The paper page lets people discuss about your paper and lets them find artifacts about it (your models for instance) you can also claim the paper as yours which will show up on your public profile at HF.

Would you like to host the model you've pre-trained on https://huggingface.co/models? Hosting on Hugging Face will give you more visibility/enable better discoverability. We can add tags in the model cards so that people find the models easier, link it to the paper page, etc.

If you're down, leaving a guide here. If it's a PyTorch model, you can use PyTorchModelHubMixin class which adds from_pretrained and push_to_hub to the model which lets you to upload the model and people to download and use models right away.
If you do not want this and directly want to upload model through UI or however you want, people can also use hf_hub_download.

After uploaded, we can also link the models to the paper page (read here) so people can discover your model.

You can also build a demo to your model on Spaces we can provide you an A100 grant.

Kind regards,

Niels

[haofeixu]

Hi @NielsRogge ,

Thank you for reaching out! Actually I have planned to use huggingface to host my models before receiving your message, as what I have done for my previous projects before :) (https://huggingface.co/spaces/haofeixu/unimatch and https://huggingface.co/haofeixu/murf).

Also, I have uploaded the DepthSplat models on huggingface: https://huggingface.co/haofeixu/depthsplat

Thank you again for creating such a great platform!

Best,
Haofei

[NielsRogge]

Hi @haofeixu, thanks a lot for uploading your model. 🤗

Regarding https://huggingface.co/spaces/haofeixu/unimatch => the demo looks great! Would be nice to remove the pre-trained weights from the Space though, and instead push them to a dedicated model repository, which is linked to the paper in similar fashion. If you then leverage hf_hub_download inside the Space, the model and Space repositories will be automatically linked.

https://huggingface.co/papers/2312.04565 looks great too, as the model is linked!

[NielsRogge]

I looked a bit into the UniMatch code base and it's a good use case for the PyTorchModelHubMixin.

I opened autonomousvision/unimatch#65, would be great if you could take a look :)